Identification and Prognosis of Patients With Interstitial Pneumonia With Autoimmune Features

Nikhil Jiwrajka, MD; Giorgos Loizidis, MD; Karen C. Patterson, MD; Maryl E. Kreider, MD, MSCE; Cheilonda R. Johnson, MD, MHS; Wallace T. Miller, Jr, MD; Eduardo Jose Mortani Barbosa, Jr, MD; Namrata Patel, MD; Michael F. Beers, MD; Leslie A. Litzky, MD; Michael D. George, MD, MSCE; Mary K. Porteous, MD, MSCE


J Clin Rheumatol. 2022;28(5):257-264. 

In This Article

Materials and Methods

Development of the Penn ILD Registry

This study was conducted at the Hospital of the University of Pennsylvania, a tertiary referral center. New and returning patients seen in the pulmonary ILD clinic between July 1, 2013, and December 1, 2017 (n = 659 total), were evaluated by a pulmonologist with expertise in ILD and consecutively enrolled in the Penn ILD Registry. The ILD diagnosis for all 659 patients was established by a multidisciplinary discussion (MDD) including at least 2 pulmonologists (one of whom is the patient's pulmonologist), a pulmonary pathologist, and a thoracic radiologist, based on a review of all of the available clinical information, in accordance with ATS/European Respiratory Society diagnostic criteria.[14–17] Rheumatologists were not included in the MDD. Demographic, clinical, functional, radiographic, serologic, histologic, and therapeutic features at baseline and follow-up ILD clinic visits were systematically recorded in the Penn ILD Registry at the time of the MDD. The creation of the registry was approved by the University of Pennsylvania's Institutional Review Board (Protocol 817680).

Study Design and Data Collection

In late 2019, we performed a medical record review of 456 patients in the Penn ILD Registry with MDD diagnoses of IPF, CTD-ILD, idiopathic nonspecific interstitial pneumonia (NSIP), cryptogenic organizing pneumonia, lymphoid interstitial pneumonia, or unclassifiable disease, excluding the 203 patients with other MDD diagnoses because the presence of an alternative ILD diagnosis would preclude IPAF, IPF, and CTD-ILD classification. Because IPAF was not an available diagnostic category during the MDD review, we retrospectively applied IPAF criteria to the population of 456 patients and used baseline and MDD data to classify 299 patients as IPAF, CTD-ILD, or IPF (see definitions below). We excluded the remaining 157 patients with ILD because of a suspected alternate or a persistently unclassifiable cause.

The baseline visit, which was defined as the patient's first clinic visit in the pulmonary ILD clinic, was used to extract demographic and clinical data and also confirm clinical features and other diagnostic data. Clinical characteristics, serologies, and other diagnostic evaluations documented within 6 months of the baseline ILD clinic visit were considered present at "baseline" to enable completion of the initial diagnostic evaluation and to avoid premature classification. The date of ILD diagnosis was specified as the date of the first computed tomography (CT) scan demonstrating ILD. Follow-up visits with the ILD clinic and, when applicable, with rheumatology were reviewed to identify new clinical symptoms/signs, serologies, radiographic patterns, histopathology, medications, and, when applicable, the type and date of CTD diagnosis. Most serologies in this study were performed by our institution's laboratory. Myositis panels were most commonly performed by ARUP Laboratories (Salt Lake City, UT).

Definitions and Classification Criteria

Patients were classified as IPAF according to previously described classification criteria (Supplementary Figure 1,[5] Briefly, patients demonstrated chest CT or histopathologic evidence of ILD that was not attributable to another cause, including CTD. Patients then fulfilled the established criteria for at least 2 of the 3 IPAF domains: clinical, serologic, and morphologic. Criteria used to identify patients with morphologic domain features, which are not explicitly defined in the IPAF classification criteria, are described in Supplementary Methods 1 (

Patients were classified as CTD-ILD based on the American College of Rheumatology classification criteria for a CTD.[18–23] When this information was unavailable, we assigned the diagnosis of CTD-ILD if a patient's rheumatologist had documented the diagnosis in at least 2 consecutive encounters and initiated disease-directed therapy in the setting of supportive clinical and diagnostic features. Specific criteria for CTDs whose classification criteria may overlap with those of IPAF, including MCTD, antisynthetase syndrome (ASyS), and anti-MDA5 dermatomyositis, are described in Supplementary Methods 2 (

Finally, patients were classified as IPF based on the initial MDD diagnosis and the presence of usual interstitial pneumonia (UIP) in the absence of an alternate explanation or sufficient IPAF criteria (i.e., truly idiopathic UIP). Usual interstitial pneumonia was defined based on the presence of "definite" or "possible" UIP upon multidisciplinary review of the high-resolution CT (HRCT) scan or the presence of UIP on histopathology regardless of the HRCT classification.[15,24]

Follow-up data unavailable at the time of the initial MDD, such as surgical pathology, were used to update clinical data for MDD ILD diagnoses (Supplementary Methods 3,[25]

Statistical and Transplant-free Survival Analysis

We performed pairwise comparisons of baseline characteristics, comparing patients with IPAF versus patients with CTD-ILD and patients with IPAF versus patients with IPF using equal-variance 2-tailed Student t tests for normally distributed continuous variables, Wilcoxon rank sum tests for non–normally distributed continuous variables, and χ 2 or Fisher exact test for categorical variables. In order to standardize the duration of follow-up after baseline visit across all 3 groups and minimize selection bias in our analysis, we restricted our transplant-free survival analysis to only those patients whose first ILD clinic visit occurred after 2010. We compared transplant-free survival across groups using Kaplan-Meier survival analysis and log-rank tests. We censored patients 1 year after the last clinic visit if they were lost to follow-up, as nearly all patients in the Penn ILD Registry would be seen in routine follow-up within a 1-year interval, or on November 30, 2019, whichever occurred sooner. For patients who died without an available date of death (n = 8 of 72 deaths; n = 1 IPAF, n = 1 CTD-ILD, n = 6 IPF), we specified a date of death 3 months (the median time between the last clinic visit and death for patients with a known date of death) from the last visit date. We also conducted a sensitivity analysis in which CTD-ILD patients with ASyS without idiopathic inflammatory myopathy were classified as IPAF to enable comparisons with previous studies. We compared transplant-free survival in age- and sex-adjusted Cox proportional hazards regression models.

Prognostic features associated with transplant-free survival in IPAF patients were assessed using univariate Cox proportional hazards regression models, with significant factors identified also evaluated in multivariable Cox models adjusted for baseline age and forced vital capacity (FVC). Statistical analyses were performed using Stata 15.1. p < 0.05 was considered statistically significant.